- Top of page
- Material and methods
Airborne laboratory-animal allergens can be measured by several methods, but little is known about the effects of important differences in methodology. Therefore, methods used in research projects in The Netherlands, the UK, and Sweden were compared. Seventy-four sets of three parallel inhalable dust samples were taken by a single operator in animal facilities in the three countries, and analyzed in parallel by the three institutes for rat and mouse urinary allergen. Rat-allergen levels measured by RAST inhibition (UK) were 3000 and 1700 times higher than levels measured by enzyme immunoassay (EIA)-sandwich methods with polyclonal rabbit (The Netherlands) or monoclonal mouse (Sweden) antibodies, while the difference between the two EIA-sandwich methods was much smaller: a factor of 2.2. For mouse allergen, an inhibition radioimmunoassay (RIA) with rabbit antimouse antibodies (UK) gave 4.6 and 5.9 times higher concentrations than sandwich EIAs with rabbit polyclonal antibodies (Sweden and The Netherlands), while the difference between the two sandwich EIAs was, on average, 1.6-fold. Thus, although levels of rat and mouse aeroallergens are significantly correlated, the assay type gives large differences in absolute concentrations, and interlaboratory technical differences affect even the same assay type. Conversion factors can aid comparison between studies, and, in the long term, assay standardization is desirable.
Abbreviations: NHLI: National Heart and Lung Institute (UK); WAU: Wageningen Agricultural University (The Netherlands); NIWL: National Institute for Working Life (Sweden); EIA: enzyme immunoassay; RAST: radioallergosorbent test; RIA: radioimmunoassay; MUA: mouse urinary allergen; RUA: rat urinary allergen; BSA: bovine serum albumin; HSA: human serum albumin; PBS: phosphate-buffered saline; PTFE: polytetrafluoroethylene (Teflon); CI: confidence interval.
Laboratory animal workers are at high risk of developing laboratory-animal allergy (LAA) (1, 2). The risk increases with the level of allergen exposure as measured by immunoassays (3–5), which can also be used to identify determinants of laboratory-animal allergen exposure (6–15). Yet, allergen concentrations should be compared with care ( 16). The reported allergen levels may differ for methodological reasons: sampling equipment, extraction methods, reference allergens, antibodies, and the design of the immunoassay used.
Therefore, we compared methods to measure rat urinary aeroallergens (RUA) (13, 17, 18) and mouse urinary aeroallergens (MUA) (18–20) in inhalable dust samples as part of the concerted action program, “Epidemiology of occupational allergic asthma and exposure to bioaerosols”, supported by the European Union. These methods have been applied in epidemiologic and other studies by three European research groups at the National Heart and Lung Institute, London, UK (NHLI) (4, 9–12, 16, 21); Wageningen Agricultural University (WAU), The Netherlands (5, 18); and the National Institute for Working Life, Solna, Sweden (NIWL) (13, 16, 20).
Material and methods
- Top of page
- Material and methods
Parallel ambient air inhalable dust samples were taken in The Netherlands (3×35=105), the UK (3×14=42), and Sweden (3×25=75) ( Fig. 1). Samples were taken in triplicate in rat rooms, mouse rooms, and cage-cleaning rooms. Additionally, 3×18 (54) blank filters were collected by mounting them in sampling heads, packing them in the same way as the test filters, and taking them to the sites of sampling, but not unpacking them. Sampling was performed by one operator (J.T.) using portable pumps at 2 l/min airflow, following in each country the current method of the participating institution ( Table 1). For each set, the three sampling heads were randomly placed approximately 15 cm apart on the sampling stand. Sampling time varied between 40 min and 20 h (mean 460 min) to provide a sufficiently wide range of RUA and MUA concentrations. Each institution received one filter from each set (74 test samples plus 18 blanks=92), without knowing its type and origin, and eluted and assayed them within 3–6 months after sampling for both rat and mouse allergens, following its published procedures (13, 17–20) (Table 1).
Table 1. Essential features of and differences between methods developed by three institutions to measure airborne rat and mouse urinary allergen levels
|Sampling method|| || || |
|Inhalable dust sampler||Seven-hole||IOM||IOM|
|PTFE filter, pore size||1.2 μm||1.0 μm||1.0 μm|
|Elution method|| || || |
|Buffer||2 ml 0.1 M NH4HCO3 plus 0.5% Tween 20||2 ml 0.15 M PBS||1 ml 0.15 M PBS plus 0.5% Tween 20|
|Method (extracts were all stored at −20°C)||Vortexed, and after 2 h, centrifuged and lyophilized. Reconstituted in PBS plus 0.3% w/v HSA before assay to get 10-fold concentrated extract ||Vortexed 2 min, sonicated 2 min, vortexed 5 min, sonicated 2 min, centrifuged ||Rotation 1 h, filter discarded, and 1% w/v BSA added|
|Immunoassay||RAST inhibition||EIA sandwich||EIA sandwich|
|Rat urine standard||From male, postpubertal Wistar rats||From young/old and male/female Wistar rats||Rat n 1.02 from 3–4-month-old male Sprague Dawley rats|
|Antibodies||IgE pool of eight rat-allergic workers||Polyclonal antibodies against RUA||Monoclonal antibodies against Rat n 1.02|
|Detection limit assay||50 ng dry weight (16.5 ng protein)/ml||0.075 ng protein/ml||0.10 ng protein/ml (unamplified protocol)|
|Detection limit method*||10 ng per filter (10.9 ng/m3) ||0.15 ng per filter (0.16 ng/m3) ||0.10 ng per filter (0.11 ng/m3) |
| ||(3.3 ng per filter [3.6 ng/m3] as protein) || || |
|Specificity||Rat urine allergens||Rat urine proteins||Rat n 1.02|
|Reproducibility||Interassay CV** 7.0%||Interassay CV** 12.9%||–|
|Immunoassay||Competitive inhibition RIA||EIA sandwich||EIA sandwich|
|Mouse urine standard||From male, postpubertal mice||From young/old and male/female Balb/c mice||Mus m 1 from postpubertal male NMRI mice|
|Antibodies||Polyclonal antibodies against MUA||Polyclonal antibodies against MUA||Polyclonal antibodies against Mus m 1|
|Detection limit assay||0.5 ng dry weight (0.09 ng protein)/ml||0.075 ng protein/ml||0.10 ng protein/ml|
|Detection limit method*†||4.0 ng per filter (4.3 ng/m3) ||0.15 ng per filter (0.16 ng/m3) ||0.10 ng per filter (0.11 ng/m3) |
| ||(0.72 ng per filter [0.8 ng/m3] as protein) || || |
|Specificity||Mouse urine proteins||Mouse urine proteins||Mus m 1|
|Reproducibility||Interassay CV** 20%||Interassay CV** 12.9%||–|
The separate effects of elution method and immunoassay were studied by exchanging filter extracts produced at WAU (n=74) and NIWL (n=74), which were reanalyzed approximately 9 months after the first analysis. Both institutions simultaneously reanalyzed the retained extracts to control for insufficient reproducibility or effects of more prolonged storage.
The detection limit of each method has been described elsewhere (13, 17–20) (Table 1). These limits and the mean sampled volume (0.92 m3) were used to calculate detection limits per m3. Statistical analyses were performed with SAS software (version 6.09). Median RUA and MUA levels were calculated, because distributions were not clearly normal or log-normal. Agreement between methods was determined by calculating the geometric mean of the ratio among allergen levels detected by different methods in parallel samples ( 22), in which both methods gave detectable levels, and a 95% confidence interval (CI) for the mean ratio was calculated for each comparison.
- Top of page
- Material and methods
At present, there is no standard method for measuring laboratory animal allergen in airborne dust samples. The present study aimed to compare methods currently available for the measurement of airborne rat and mouse allergens. The methods had been independently developed for specific studies by three research groups and used to investigate the exposure–response relationship between exposure to allergen and the development of LAA. This study showed that results obtained by different methods give similar relative values but that absolute concentrations showed large differences. Some of the main determinants causing these differences could be identified, while a more detailed analysis is presented in an accompanying paper ( 23).
In theory, differences could be due to dust-sampling method, filter elution, storage in the laboratory, or immunoassay. This study did not attempt to evaluate the effect of dust sampling because this has been studied by others (24, 25). All three institutions sampled the inhalable dust fraction. Although they did use different makes of sampling head, this is not estimated to make more than a 1.2-fold difference in total dust collected. However, sampling heads designed to collect different dust fractions would be expected to give large differences.
It has been shown previously that the addition of detergent increases the recovery of rat urinary allergen from PTFE and glass-fiber filters ( 26), especially at low levels of allergen, when most of the allergen load is likely to be deeply embedded in the filter. There have been similar findings for grass pollen ( 27) and for other biologically relevant components such as endotoxins, the size of the effect probably depending on solubility (28–30). In the present study, the elution factors which varied were use of a detergent, amount of agitation, and duration of elution. The filter-exchange study by NIWL and WAU confirmed that the use of a detergent increases allergen yield by one order of magnitude. NHLI did not join in this study but used detergent and had the longest duration of elution (Table 1).
Extensive agitation of PTFE filters is not needed for the efficient elution of rat urinary allergen ( 26). Furthermore, rat urinary allergens may not be stable if sonicated (S. Gordon, unpublished observation); therefore, prolonged sonication for 4 min, as used by WAU, especially in the absence of added protein, may destroy allergen and reduce the yield. Another factor explaining the low yields at WAU is the halving in allergen yield after storage of frozen extracts for approximately 9 months. WAU, unlike NHLI and NIWL, did not store its extracts with added detergent or protein, which may prevent adherence of proteins to vial walls and pipette tips.
An important finding of this study is that air samples need to be processed in the same way before immunoassay in order to make meaningful comparisons between air samples taken at different times or in different working areas. Otherwise, erroneous conclusions could be drawn about, for example, the effectiveness of control measures. The variations due to elution and storage methods discussed above were of a magnitude sufficient to explain most of the differences observed between the MUA levels measured by the three centers. The MUA immunoassays were similar, each employing rabbit polyclonal antibody, but there may be a small assay effect, which is discussed further in the companion paper ( 23).
Elution and storage effects do not explain all the differences in RUA levels, and the remaining differences must be attributable to differences in immunoassay. This is explored further in the companion paper ( 23). When elution and storage effects were controlled, the monoclonal immunoassay (NIWL) detected less RUA than the rabbit polyclonal immunoassay (WAU). This is not surprising, since a monoclonal assay is designed to detect only one component, in this case, the major allergen Rat n 1.02 ( 13), whereas rat-room air is complex, containing many major and minor allergens. Assays using experimentally induced animal polyclonal antibodies (as at WAU) or human polyclonal antibodies from occupational sensitization (as at NHLI) would be expected to capture a greater range of antibodies. It is of interest that the greater yield of allergen from filters by NIWL (and NHLI) methods appeared to outweigh the narrow specificity of the monoclonal technology, because the RUA levels measured by NIWL in parallel filters were consistently higher than those reported by WAU (Fig. 2c).
There are also arithmetic factors which differ between the immunoassays. Both the WAU and NIWL results were expressed in relation to the protein content of the standard employed. However, the NHLI results were not corrected for protein and were simply expressed as dry weight. This is primarily because carbohydrate or glycosylated allergens in the dust or standard urine extract may contribute to the inhibition seen in the UK assays, but also because the protein content of dust samples is usually much lower than that of urine ( 31). As the protein content of the NHLI RUA and MUA extracts was 35.4% and 17.0%, respectively, the NHLI values should be reduced by multiplying by 0.35 and 0.17 if comparison of the protein content of the filters is required.
Allergen was detected on blank filters by each method, most frequently in the NHLI assays. This phenomenon has been noted before, particularly for mouse ( 19). However, the positive values for blank filters were consistently some orders of magnitude lower than for test filters. Since these filters were taken into the animal-house environment but not exposed, this effect may reflect minor contamination of sampling heads or filters. Alternatively, these may be false positives due to inherent problems in the detection of low levels of allergen caused by a shallow standard curve at low concentrations.
The assays were, as expected, immunologically specific, detecting RUA in rat rooms and MUA in mouse rooms. All three systems also detected RUA in mouse rooms and vice versa, although at much lower levels than in the homologous rooms (Table 3). The NHLI assay had a greater number of nonhomologous positives. The most likely explanations are residual allergen from prior occupation of the room by the other species, and contamination through doors, on staff clothing, or via ventilation. Contamination of sampling heads or in the analyzing laboratories is less likely, since that should have equally affected the results for the blank filters. Cross-reaction between species in all assays was minimal ( 23).
In the absence of a “gold standard”, centers will continue to use the assay which appears best for their purposes. There are reasons not addressed in this study for choosing one assay rather than another. Monoclonal technology, as used by NIWL for the RUA assay, offers good reproducibility, so that comparisons can be made at very different time periods. Techniques using human IgE pools, as used by NHLI in the project contributing to this study, can be expected to measure allergens of clinical relevance, while assays using other antibody sources always require further validation. These issues aside, this comparative study has shown that identical air samples would give different results depending on the assay technology employed.
These differences are by orders of magnitude and would clearly be important if quantitative comparisons between workplaces or over time were to be made without taking account of sample processing, storage, assay, and arithmetic assumptions. The three methods, however, gave rankings similar to the air samples (Figs. 2 and 3), so that internal comparisons within studies are valid whatever assay is used and semiquantitative approaches – for example, using percentiles of the exposure distribution – may also be valid. In principle, the use of conversion factors, obtainable, for example, from the results of this study, could make data from allergen measurements comparable between studies. For future work, however, standardization of assays is desirable.